In India where 65.53 per cent population resides in villages, as per a World Bank report of 2019, there is a significant digital divide that exists. Hence, there is a dire need to develop technologies accessible to all and operable by all. Addressing TECH5’s initiatives of building AI-based and touchless fingerprint technologies that are inclusive and causing innovation in government, Rahul Parthe, Chief Technology Officer, Co-Founder and Chairman, TECH5, interacted with Harshal Desai of Elets News Network (ENN).

Please inform us about the operations of TECH5

Let me begin with the name TECH5, which is an abbreviation of Tech FI, or “Technology for Inclusion.” Inclusion is our core principle, and for us, it means that everyone in society should be able to participate in programs that governments have designed for all people. Our technology has been created for inclusion. We have invested and built biometric technologies such as facial recognition, fingerprint recognition, and iris recognition, which allow us to ensure that every available program can be made accessible for every citizen. Aadhaar, a very good example of such a project, implemented all three modalities. In the output of the design phase, the decision was made as to which biometric modality to use.

We, as TECH5, are purely technology providers, and our goal is to build the latest and greatest biometric technologies, such as those used in Aadhaar. However, at the same time, we are also focused on building disruptive and innovative solutions for credentialing. These technologies will help to ensure the issuing of IDs that everybody in society can use. As we all know, there is a digital divide that has been made more evident and has widened as a result of COVID-19, we make sure that we build technologies that every person in society can access and leverage in order to improve their standard of living.

Could you also brief us about innovations that TECH5 is working on and how R&D has evolved within TECH5 to ensure continuity in bringing innovations to the industry?

TECH5 is a technology provider founded by people who have extensive experience in the field of biometrics and identity management. I was personally involved in Aadhaar from day one, so I have seen how it has grown from nothing to what it is today. My co-founder and other employees come from similar backgrounds and experiences. We are taking the vision of making technology available for everybody very seriously, and are continuously investing in building and improving these technologies.

As you may have heard, there is a new revolution called AI and Machine Learning, and we are part of this revolution. We have taken the concepts of AI and Machine Learning and applied them to biometric algorithm development. The benefit has been obvious. Technologies driven by Machine Learning have outperformed traditional approaches and we have achieved exponential growth in accuracy and performance in facial recognition technology. It is now the basic principle that motivates our approach for our other technologies, such as fingerprint and iris recognition. We have already achieved a world top-tier ranking in facial recognition in NIST (National Institute of Standards and Technology) – the most well known international benchmark in the industry, where you prove how well your technology performs. Based on that success, we applied our expertise in that approach to fingerprint. We are happy to confirm that we have already achieved groundbreaking results in our fingerprint matching technology that is in line with our face recognition technology.

The impact of applying AI to fingerprint recognition has resulted in much-improved accuracy as well as a reduction of the total cost of ownership. Imagine if today Aadhaar could spend less than a fifth of its budget on hardware that is currently required for data centres where biometric matching is performed. I am sure, with these advancements, TECH5 will soon be able to stake the claim as the world’s fastest and most accurate fingerprint matching technology in addition to our already successful facial recognition technology. We are currently also working on our iris recognition technology to make it a top-performing solution that we can offer to our customers and business partners.

Fingerprint recognition tends to make errors. How can such issues be addressed to ensure accuracy?

This issue also can be solved by applying AI and Machine Learning. For traditional algorithms, an engineered model has usually been used. This basically means that humans thought about how fingerprints looked and then came up with an algorithm that compared fingerprint features. But, the problem is that if something changes, even by a small margin, the algorithm starts to underperform. Machine Learning is like a human brain. For example, if a person shaves, puts on a new pair of glasses or something similar that covers half of the face, both the human brain and Machine Learning will still be able to accurately recognise that person.

There is, nevertheless, a fundamental limit. If, for example, I undergo plastic surgery and start to look like Tom Cruise, nobody in the world will recognise me, and the same is true for all biometrics. On the other hand, an event like ageing will not stop the technology from accurately recognising a person. Our mission is to make sure that the technology works well in non-ideal conditions – in virtual darkness, for example. It is a race to zero error.

How touchless technologies for fingerprint recognition are different from classic software-hardware solutions, and how the new technology can revolutionize finger capturing, enrolment and recognition processes?

As we all know, some time ago fingerprints were captured using ink and paper. This is how it all began. Then, a digital form of fingerprint scanning called live scan was invented. I’m sure most of us have already had some personal experience with this. For example, a single fingerprint scanner is used for opening a door, and a 4-4-2 fingerprint scanner can then be used for enrollment in a National ID project like Aadhaar. This approach involves two components: one is a dedicated piece of hardware that is connected to a computer, phone, or laptop, and is effectively a touch solution – you have to touch the scanner, and the second is a software suite. As a result, there are a lot of logistical issues that have to be faced to make sure that the devices are available because people have to interact with the device. In addition, you also have the traditional issues of supply chain management and supporting the infrastructure.

In the meantime, mobile phone technology has been developing exponentially, not only in terms of penetration as almost everybody now has a phone but also in terms of its functionality. Today’s smartphone is as powerful as a laptop was a decade ago, and it not only has very high processing power but also comes with integrated high-end cameras, sometimes four or five cameras on a single phone, with very high resolution, flash, autofocus and other features. We took into consideration the fact that the previous challenges in biometric capture were caused by connected devices and infrastructure, as well as the touch limitation, and took the approach of leveraging the powerful smartphone device that you carry with you to capture fingerprints. And, as you can imagine, it was a no-brainer, because the capability of capturing fingerprints using a mobile phone translated into a significant increase in the range of possibilities. Anyone, anywhere as a part of a particular program can take a medium-range mobile phone, which is very common today, and capture one’s fingerprints within seconds. This achievement has solved the problem of distribution and has provided fast access to the technology that helps to avoid any kind of touch, a requirement which has become extremely important during the pandemic.

How will contactless technologies perform globally in order to continue to penetrate business, even in times of the pandemic?

The move towards touchless technology started before the pandemic and we were trying our best to explain to people why the technology was important. Unfortunately, COVID-19 happened, and now the value of touchless technologies has become very obvious to everyone. In terms of TECH5 biometric technologies, both face and iris are already touchless, and until recently only fingerprint recognition required touch. Now, we have made fingerprint recognition technology touchless as well. This touchless technology not only helps liberate the capture technology, making it accessible for everybody but also makes readily available a method for capturing biometrics safely during this or any other emergency that would require no contact.

It is important to note that this technology does not impact any of the regulatory or security policies. What we are doing is only changing a scanner to a touchless technology and saying, “Let’s use the revolutionary AI and Machine Learning-based matching algorithm.” Also important to note is that our solution works with old and new technology; they are basically interoperable. The touchless fingerprint technology can be made available to everyone while following security and regulatory policies that ensure that people’s data are securely handled, and privacy and security are not breached.

We’ve talked briefly about AI, ML and so on. How can such technologies help to leverage the entire contactless infrastructure, in your opinion? And what are the challenges that you see in this approach?

This approach has more advantages than disadvantages. As I previously said, the problem needs to be solved from two ends. Namely, it is not only about improving the capture technology but also about accurately matching the captured fingerprints, while also harmoniously working with traditional systems. We are developing these technologies using Machine Learning and AI, and they can work with existing data while also adopting new emerging concepts not only for capture but also for other facets of IT. For example, if tomorrow Aadhaar decided to go completely contactless, it would not mean that the existing data of 1.3 billion people would have to be discarded.

You have probably seen these futuristic videos (by the way, they are not so futuristic anymore), where you walk into a grocery store, pick up your groceries, and walk out. You will not have to do any touch contact or payment, and everything will be handled fairly seamlessly. This is where I see the combination of cloud infrastructure, touchless fingerprint capture and accurate matching technologies to make sure that your privacy and security is not breached and the experience is completely frictionless.

Which are the countries that you will be focusing a lot on in 5 years time?

TECH5 is a global company, and we have customers right from Latin America all the way to the far East. India is one of the focus areas for us, and we have a big team in that country. We have great natural pride in India because one of the co-founders is Indian. With respect to verticals, we are in the government sector, as well as in the private sector. We also serve law enforcement and the military. We want to have a presence anywhere, where you can use biometrics for improving both the lives of citizens and security. Our customers are mostly governments, banks, and FinTech (we provide EKYC solutions). You can see our face recognition and fingerprint solutions at work across different industries. We are there to serve our partners.

What is TECH5’s plan for further development of its technology?

As I previously mentioned, we take the company’s vision of making technologies available and inclusive very seriously, so we continue to invest in this. We also believe that the potential in AI and Machine Learning has not been fully realized. As such, we are not only focusing on achieving greater results in capturing and matching fingerprints but have also developed a new concept called T5-Mobile ID that can replace Smart card-based solutions and eliminate the need for carrying a physical ID. Everything an individual will need, for example, a health care card, a bus ticket, or a bank card will be available in electronic form. The solution is entirely electronic and as a result, the issuer will not need to spend money on chips or cards, giving the solution the added advantage of being environmentally friendly.

All of the solutions that we will continue to build will be meant for every person in the society, and not only for the elite or privileged few. We want to make sure that people with and without digital skills will be able to use our solutions and thereby contribute to bridging the digital gap. To reach our goals, we are continuing to invest in all of the components of the solution that are required to ensure that identity solutions are acceptable, available, frictionless, and private for everyone in society.

Are we making the most of touchless technology in India? Are there sectors that are lagging behind?

I believe that every sector that involves fingerprints will be able to use this technology. We are experimenting and running it with customers in different parts of the world, in various use cases. For example, we have law enforcement customers who have used it for a criminal investigation to find a latent print of a crime scene. It has also been used as a part of eKYC for opening a bank account using face or fingerprint recognition technology. The last point that I would like to make is that we are designing disruptive technologies that work in non-ideal conditions to make sure that both capture and remote one-to-one matching based on T5-Mobile ID technology can be delivered even in very remote usually unserved areas without having a need to connect to any backend system.

Any message that you would like to give to the industry, your partners, and your current and potential customers?

I would like to conclude by saying that TECH5 is a company founded by veterans who have worked in this field for decades. We believe in the revolution that AI is spawning and have already invested in it and proven that with these advancements we will have very good results in a very short period of time. We are working towards educating our partners and customers so that they start looking at these new approaches, not only from the cost-saving perspective but also in terms of the enhancement in accuracy and user experience. At the end of the day, the question is really about whom you are serving. As we build relationships with our partners and customers, the ultimate goal is for users (citizens) to be happy, to like the experience, and do not have to worry about their security and privacy. If the goal is achieved, then everyone benefits. Finally, we would like to take this opportunity to say loud and clear that we are a top player and an expert in the industry, and we are here and happy to talk to anyone who is interested in these advancements in biometric technologies.

Link to media coverage